Bug 68735

Summary: Cant boot 5.2.1-RELEASE GENERIC kernel
Product: Base System Reporter: Adam Retter <adam>
Component: kernAssignee: Søren Schmidt <sos>
Status: Closed FIXED    
Severity: Affects Only Me    
Priority: Normal    
Version: 5.2.1-RELEASE   
Hardware: Any   
OS: Any   

Description Adam Retter 2004-07-06 20:10:24 UTC
I have been trying to cvsup from 5.2-RELEASE to 5.2.1-RELEASE. 5.2-RELEASE had been installed for severall weeks and working fine.

I can boot the 5.2-RELEASE GENERIC kernel, but if I build the 5.2.1-RELEASE GENERIC kernel after a cvsup of /usr/src and then a buildworld and reboot the machine. The system hangs whilst booting the 5.2.1-RELEASE GENERIC kernel with the following messages -

Mounting Root from ufs: /dev/ad8s2a
ad8: TIMEOUT READ_DMA retrying (2 retries left) LBA=51199379
GEOM: destroy disk ad8 dp=0xc6a4ea60
ad8: WARNING - removed from configuration

From http://www.uk.freebsd.org/releases/5.2.1R/announce.html I see this "Significantly improved ATA/IDE and SATA handling. Problems with timeouts, error recovery, and certain master/slave configurations have been largely addressed."

Unfortunately these significant improvements have actually broken my SATA support :-( This is most upsetting as my SATA drives wouldnt work under 4.x, 5 or 5.1 but did work nicely under 5.2, but 5.2.1 seems to be a step backwards.

I have an MSI 848P-Neo Motherboard that uses the Intel ICH5 SATA Chipset and I am using two Maxtor SATA hard-disks (1x120GB and 1x160GB).

Can someone help me or can I help someone debug this problem and get my SATA support working again, I really dont want to go to -CURRENT as this is my workstation machine, I usually use -RELENG.

Thanks

Fix: 

Merge the SATA code for the ICH5 from 5.2 with that from 5.2.1???2
How-To-Repeat: 
Get an MSI 848P-Neo motherboard and a SATA hard-disk and try and boot 5.2.1-RELEASE.
Comment 1 Tilman Keskinoz freebsd_committer freebsd_triage 2004-08-24 20:08:11 UTC
Responsible Changed
From-To: freebsd-bugs->sos

Another PR for the ATA maintainer.
Comment 2 Søren Schmidt freebsd_committer freebsd_triage 2004-08-24 20:58:50 UTC
State Changed
From-To: open->closed

The ich5 problems has been solved in -current fro quite some time.